
import pandas as pd
import numpy as np

df = pd.read_csv('earthquakes_2023_global_bruite.csv', delimiter = ';')

df


df = df.drop_duplicates()
df


df = df.dropna()
df


df.time

0        2023-01-01T00:49:25.294Z
1        2023-01-01T01:41:43.755Z
2        2023-01-01T03:29:31.070Z
4        2023-01-01T04:29:13.793Z
5        2023-01-01T04:50:17.639Z
                   ...           
26637    2023-12-29T03:37:19.334Z
26638    2023-12-29T04:38:54.109Z
26639    2023-12-29T08:42:05.747Z
26640    2023-12-29T11:02:48.679Z
26641    2023-12-29T16:31:16.656Z
Name: time, Length: 21080, dtype: object


df.info()

<class 'pandas.core.frame.DataFrame'>
Index: 21080 entries, 0 to 26641
Data columns (total 22 columns):
 #   Column           Non-Null Count  Dtype  
---  ------           --------------  -----  
 0   time             21080 non-null  object 
 1   latitude         21080 non-null  float64
 2   longitude        21080 non-null  float64
 3   depth            21080 non-null  float64
 4   mag              21080 non-null  float64
 5   magType          21080 non-null  object 
 6   nst              21080 non-null  float64
 7   gap              21080 non-null  float64
 8   dmin             21080 non-null  float64
 9   rms              21080 non-null  float64
 10  net              21080 non-null  object 
 11  id               21080 non-null  object 
 12  updated          21080 non-null  object 
 13  place            21080 non-null  object 
 14  type             21080 non-null  object 
 15  horizontalError  21080 non-null  float64
 16  depthError       21080 non-null  float64
 17  magError         21080 non-null  float64
 18  magNst           21080 non-null  float64
 19  status           21080 non-null  object 
 20  locationSource   21080 non-null  object 
 21  magSource        21080 non-null  object 
dtypes: float64(12), object(10)
memory usage: 3.7+ MB


df['time'] = pd.to_datetime(df['time'], format='%Y-%m-%dT%H:%M:%S.%fZ')

df['updated'] = pd.to_datetime(df['updated'], format='%Y-%m-%dT%H:%M:%S.%fZ')

df['time']

df['updated']

C:\Users\Killian\AppData\Local\Temp\ipykernel_20516\1747763810.py:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df['time'] = pd.to_datetime(df['time'], format='%Y-%m-%dT%H:%M:%S.%fZ')
C:\Users\Killian\AppData\Local\Temp\ipykernel_20516\1747763810.py:3: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df['updated'] = pd.to_datetime(df['updated'], format='%Y-%m-%dT%H:%M:%S.%fZ')

0       2023-03-11 22:51:52.040
1       2023-03-11 22:51:45.040
2       2023-03-11 22:51:29.040
4       2023-03-11 22:51:38.040
5       2023-03-11 22:51:29.040
                  ...          
26637   2023-12-29 04:05:57.040
26638   2023-12-29 10:59:44.533
26639   2023-12-29 08:57:05.040
26640   2023-12-29 11:22:46.040
26641   2023-12-29 16:45:27.040
Name: updated, Length: 21080, dtype: datetime64[ns]


from sklearn.preprocessing import LabelEncoder

label_encoder = LabelEncoder()

colonnes_a_exclure = ['place', 'id', 'type']

for colonne in df.columns:
    if colonne not in colonnes_a_exclure and df[colonne].dtypes == 'object':
        df[colonne] = label_encoder.fit_transform(df[colonne])
        print(df[colonne])

df.info()

0        3
1        0
2        2
4        3
5        3
        ..
26637    0
26638    0
26639    0
26640    0
26641    0
Name: magType, Length: 21080, dtype: int32
0        7
1        7
2        4
4        7
5        7
        ..
26637    7
26638    7
26639    7
26640    7
26641    7
Name: net, Length: 21080, dtype: int32
0        1
1        1
2        1
4        1
5        1
        ..
26637    1
26638    1
26639    1
26640    1
26641    1
Name: status, Length: 21080, dtype: int32
0        7
1        7
2        4
4        7
5        7
        ..
26637    7
26638    7
26639    7
26640    7
26641    7
Name: locationSource, Length: 21080, dtype: int32
0        7
1        7
2        4
4        7
5        7
        ..
26637    7
26638    7
26639    7
26640    7
26641    7
Name: magSource, Length: 21080, dtype: int32
<class 'pandas.core.frame.DataFrame'>
Index: 21080 entries, 0 to 26641
Data columns (total 22 columns):
 #   Column           Non-Null Count  Dtype         
---  ------           --------------  -----         
 0   time             21080 non-null  datetime64[ns]
 1   latitude         21080 non-null  float64       
 2   longitude        21080 non-null  float64       
 3   depth            21080 non-null  float64       
 4   mag              21080 non-null  float64       
 5   magType          21080 non-null  int32         
 6   nst              21080 non-null  float64       
 7   gap              21080 non-null  float64       
 8   dmin             21080 non-null  float64       
 9   rms              21080 non-null  float64       
 10  net              21080 non-null  int32         
 11  id               21080 non-null  object        
 12  updated          21080 non-null  datetime64[ns]
 13  place            21080 non-null  object        
 14  type             21080 non-null  object        
 15  horizontalError  21080 non-null  float64       
 16  depthError       21080 non-null  float64       
 17  magError         21080 non-null  float64       
 18  magNst           21080 non-null  float64       
 19  status           21080 non-null  int32         
 20  locationSource   21080 non-null  int32         
 21  magSource        21080 non-null  int32         
dtypes: datetime64[ns](2), float64(12), int32(5), object(3)
memory usage: 3.3+ MB

C:\Users\Killian\AppData\Local\Temp\ipykernel_20516\3547750565.py:9: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df[colonne] = label_encoder.fit_transform(df[colonne])
C:\Users\Killian\AppData\Local\Temp\ipykernel_20516\3547750565.py:9: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df[colonne] = label_encoder.fit_transform(df[colonne])
C:\Users\Killian\AppData\Local\Temp\ipykernel_20516\3547750565.py:9: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df[colonne] = label_encoder.fit_transform(df[colonne])
C:\Users\Killian\AppData\Local\Temp\ipykernel_20516\3547750565.py:9: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df[colonne] = label_encoder.fit_transform(df[colonne])
C:\Users\Killian\AppData\Local\Temp\ipykernel_20516\3547750565.py:9: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df[colonne] = label_encoder.fit_transform(df[colonne])


df.info()

<class 'pandas.core.frame.DataFrame'>
Index: 21080 entries, 0 to 26641
Data columns (total 22 columns):
 #   Column           Non-Null Count  Dtype         
---  ------           --------------  -----         
 0   time             21080 non-null  datetime64[ns]
 1   latitude         21080 non-null  float64       
 2   longitude        21080 non-null  float64       
 3   depth            21080 non-null  float64       
 4   mag              21080 non-null  float64       
 5   magType          21080 non-null  int32         
 6   nst              21080 non-null  float64       
 7   gap              21080 non-null  float64       
 8   dmin             21080 non-null  float64       
 9   rms              21080 non-null  float64       
 10  net              21080 non-null  int32         
 11  id               21080 non-null  object        
 12  updated          21080 non-null  datetime64[ns]
 13  place            21080 non-null  object        
 14  type             21080 non-null  object        
 15  horizontalError  21080 non-null  float64       
 16  depthError       21080 non-null  float64       
 17  magError         21080 non-null  float64       
 18  magNst           21080 non-null  float64       
 19  status           21080 non-null  int32         
 20  locationSource   21080 non-null  int32         
 21  magSource        21080 non-null  int32         
dtypes: datetime64[ns](2), float64(12), int32(5), object(3)
memory usage: 3.3+ MB


print(df.describe())

                                time      latitude     longitude  \
count                          21080  21080.000000  21080.000000   
mean   2023-06-28 14:41:49.917795840     15.680956     -3.106846   
min       2023-01-01 00:49:25.294000    -65.849700   -179.998700   
25%    2023-03-25 19:21:46.375500032     -6.513600   -118.385000   
50%    2023-06-25 20:33:20.763500032     18.227500    -26.733400   
75%    2023-10-01 13:39:05.907749888     38.858675    127.268800   
max       2023-12-29 23:17:18.800000     86.593900    179.999400   
std                              NaN     29.314531    128.808562   

              depth           mag       magType           nst           gap  \
count  21080.000000  21080.000000  21080.000000  21080.000000  21080.000000   
mean      68.321158      4.048470      1.605408     42.528795    127.081074   
min       -3.370000      2.600000      0.000000      3.000000      8.000000   
25%       10.000000      3.370000      0.000000     18.000000     73.000000   
50%       21.532000      4.300000      0.000000     29.000000    113.000000   
75%       66.333250      4.500000      3.000000     52.000000    172.000000   
max      681.238000      7.800000     11.000000    423.000000    350.000000   
std      118.155052      0.786942      2.833168     38.923838     69.076922   

               dmin           rms           net  \
count  21080.000000  21080.000000  21080.000000   
mean       2.551639      0.578899      6.551044   
min        0.000000      0.010000      0.000000   
25%        0.576975      0.400000      7.000000   
50%        1.491000      0.580000      7.000000   
75%        3.074000      0.750000      7.000000   
max       50.820000      1.880000      9.000000   
std        3.805738      0.255945      1.297837   

                             updated  horizontalError    depthError  \
count                          21080     21080.000000  21080.000000   
mean   2023-08-24 08:20:55.669221120         6.963787      4.800911   
min       2023-01-01 11:14:15.230000         0.000000      0.100000   
25%    2023-05-26 21:54:07.790000128         4.100000      1.878000   
50%       2023-08-26 20:54:24.040000         6.980000      2.693000   
75%    2023-11-28 18:44:39.960250112         9.620000      6.955250   
max       2023-12-30 00:23:54.040000        99.000000     47.878000   
std                              NaN         4.037148      4.577801   

           magError        magNst        status  locationSource     magSource  
count  21080.000000  21080.000000  21080.000000    21080.000000  21080.000000  
mean       0.118441     33.598482      0.999573        6.551044      6.551044  
min        0.000000      0.000000      0.000000        0.000000      0.000000  
25%        0.078000     10.000000      1.000000        7.000000      7.000000  
50%        0.108000     18.000000      1.000000        7.000000      7.000000  
75%        0.148000     36.000000      1.000000        7.000000      7.000000  
max        0.770000    884.000000      1.000000        9.000000      9.000000  
std        0.058194     49.353409      0.020659        1.297837      1.297837


print(df.magNst.value_counts())

magNst
10.0     1124
12.0     1012
8.0       989
14.0      933
6.0       862
         ... 
273.0       1
457.0       1
500.0       1
257.0       1
404.0       1
Name: count, Length: 379, dtype: int64


colonne_num = df.select_dtypes(include='number')
correlation_matrix = colonne_num.corr()

print(correlation_matrix)

                 latitude  longitude     depth       mag   magType       nst  \
latitude         1.000000  -0.180172 -0.279452 -0.534165  0.174312 -0.002172   
longitude       -0.180172   1.000000 -0.054961  0.491289 -0.164641  0.151647   
depth           -0.279452  -0.054961  1.000000  0.143370 -0.135157  0.064436   
mag             -0.534165   0.491289  0.143370  1.000000  0.009480  0.528265   
magType          0.174312  -0.164641 -0.135157  0.009480  1.000000  0.405277   
nst             -0.002172   0.151647  0.064436  0.528265  0.405277  1.000000   
gap              0.311072  -0.299297 -0.179294 -0.540287 -0.036322 -0.476688   
dmin            -0.355968   0.186362  0.041044  0.354928 -0.108273  0.068382   
rms             -0.219160   0.213494  0.097704  0.494904 -0.077533  0.183966   
net             -0.126499   0.233183  0.136797  0.413283 -0.090965  0.090476   
horizontalError -0.411654   0.283793  0.415469  0.470323 -0.254747  0.022009   
depthError      -0.034863  -0.034154  0.326831 -0.036556 -0.101855 -0.142591   
magError        -0.156546   0.014150 -0.036732 -0.133314 -0.277945 -0.430332   
magNst           0.027111   0.043381  0.099790  0.247017 -0.090177  0.499504   
status          -0.014022   0.018107  0.011054  0.032293 -0.009363  0.008422   
locationSource  -0.126499   0.233183  0.136797  0.413283 -0.090965  0.090476   
magSource       -0.126499   0.233183  0.136797  0.413283 -0.090965  0.090476   

                      gap      dmin       rms       net  horizontalError  \
latitude         0.311072 -0.355968 -0.219160 -0.126499        -0.411654   
longitude       -0.299297  0.186362  0.213494  0.233183         0.283793   
depth           -0.179294  0.041044  0.097704  0.136797         0.415469   
mag             -0.540287  0.354928  0.494904  0.413283         0.470323   
magType         -0.036322 -0.108273 -0.077533 -0.090965        -0.254747   
nst             -0.476688  0.068382  0.183966  0.090476         0.022009   
gap              1.000000 -0.168113 -0.262887 -0.172500        -0.130615   
dmin            -0.168113  1.000000  0.145757  0.202912         0.432698   
rms             -0.262887  0.145757  1.000000  0.486624         0.334725   
net             -0.172500  0.202912  0.486624  1.000000         0.480372   
horizontalError -0.130615  0.432698  0.334725  0.480372         1.000000   
depthError       0.239580 -0.080188  0.017430  0.000588         0.227445   
magError         0.069712 -0.010270 -0.058038 -0.075433         0.044140   
magNst          -0.296739  0.041580  0.099376 -0.015280         0.016140   
status           0.000955  0.013104  0.031942  0.054780         0.029911   
locationSource  -0.172500  0.202912  0.486624  1.000000         0.480372   
magSource       -0.172500  0.202912  0.486624  1.000000         0.480372   

                 depthError  magError    magNst    status  locationSource  \
latitude          -0.034863 -0.156546  0.027111 -0.014022       -0.126499   
longitude         -0.034154  0.014150  0.043381  0.018107        0.233183   
depth              0.326831 -0.036732  0.099790  0.011054        0.136797   
mag               -0.036556 -0.133314  0.247017  0.032293        0.413283   
magType           -0.101855 -0.277945 -0.090177 -0.009363       -0.090965   
nst               -0.142591 -0.430332  0.499504  0.008422        0.090476   
gap                0.239580  0.069712 -0.296739  0.000955       -0.172500   
dmin              -0.080188 -0.010270  0.041580  0.013104        0.202912   
rms                0.017430 -0.058038  0.099376  0.031942        0.486624   
net                0.000588 -0.075433 -0.015280  0.054780        1.000000   
horizontalError    0.227445  0.044140  0.016140  0.029911        0.480372   
depthError         1.000000  0.013334 -0.075370 -0.009574        0.000588   
magError           0.013334  1.000000 -0.427177 -0.020402       -0.075433   
magNst            -0.075370 -0.427177  1.000000  0.002391       -0.015280   
status            -0.009574 -0.020402  0.002391  1.000000        0.054780   
locationSource     0.000588 -0.075433 -0.015280  0.054780        1.000000   
magSource          0.000588 -0.075433 -0.015280  0.054780        1.000000   

                 magSource  
latitude         -0.126499  
longitude         0.233183  
depth             0.136797  
mag               0.413283  
magType          -0.090965  
nst               0.090476  
gap              -0.172500  
dmin              0.202912  
rms               0.486624  
net               1.000000  
horizontalError   0.480372  
depthError        0.000588  
magError         -0.075433  
magNst           -0.015280  
status            0.054780  
locationSource    1.000000  
magSource         1.000000


time_stat = df['time'].describe()
print(time_stat)

update_stat = df['updated'].describe()
print(update_stat)

count                            21080
mean     2023-06-28 14:41:49.917795840
min         2023-01-01 00:49:25.294000
25%      2023-03-25 19:21:46.375500032
50%      2023-06-25 20:33:20.763500032
75%      2023-10-01 13:39:05.907749888
max         2023-12-29 23:17:18.800000
Name: time, dtype: object
count                            21080
mean     2023-08-24 08:20:55.669221120
min         2023-01-01 11:14:15.230000
25%      2023-05-26 21:54:07.790000128
50%         2023-08-26 20:54:24.040000
75%      2023-11-28 18:44:39.960250112
max         2023-12-30 00:23:54.040000
Name: updated, dtype: object


import matplotlib.pyplot as plt
import seaborn as sns

plt.figure(figsize=(12, 6))
sns.boxplot(data=df)
plt.xticks(rotation=45)  
plt.show()


import plotly.express as px
earthquake_type = df['type'].value_counts()

fig = px.bar(y = earthquake_type.values, x = earthquake_type.index, color = earthquake_type.index, text = earthquake_type.values)
fig.show()


plt.figure(figsize=(10, 6))

sns.scatterplot(x='time', y='mag', data=df, s=50)

plt.xlabel('date')
plt.ylabel('magnitude')
plt.show()


import folium

carte = folium.Map(location=[43.327408, -1.032999], zoom_start = 2)
for ind, i in df.iterrows():
    if i['type'] == 'earthquake' and i['mag'] > 5:
        carte.add_child(folium.RegularPolygonMarker(location = [i['latitude'], i['longitude']], popup = i['place'], fill_color = 'red', radius = 10))

    
carte


from sklearn.model_selection import train_test_split
from sklearn.preprocessing import MinMaxScaler
from sklearn.linear_model import LinearRegression, LogisticRegression
from sklearn.neighbors import KNeighborsClassifier, KNeighborsRegressor
from sklearn.preprocessing import LabelEncoder
from sklearn.metrics import accuracy_score, confusion_matrix, mean_squared_error, classification_report, precision_score, f1_score, recall_score


X = df[['latitude', 'longitude', 'mag']]
y = (df['mag'] > 5).astype(int)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.7, random_state = 42)


scaler = MinMaxScaler()
X_train_norm = scaler.fit_transform(X_train)
X_test_norm = scaler.transform(X_test)


label_encoder = LabelEncoder()
y = label_encoder.fit_transform(y)


regression_lineaire = LinearRegression()
regression_lineaire.fit(X_train_norm, y_train)
y_pred_lineaire = regression_lineaire.predict(X_test_norm)


regression_logistique = LogisticRegression()
regression_logistique.fit(X_train_norm, y_train)
y_pred_logistique = regression_logistique.predict(X_test_norm)


knn_model = KNeighborsClassifier(n_neighbors=3)
knn_model.fit(X_train_norm, y_train)
y_pred_knn = knn_model.predict(X_test_norm)


precision_logistique = accuracy_score(y_test, y_pred_logistique)
precision_knn = accuracy_score(y_test, y_pred_knn)

print(f'Erreur quadratique (Régression Linéaire) : {mean_squared_error(y_test, y_pred_lineaire)}\n')

print(f'Précision (Régression Logistique) : {precision_logistique * 100}')
print(f'Erreur quadratique (Régression Logistique) : {mean_squared_error(y_test, y_pred_logistique)}\n')

print(f'Précision (KNN) : {precision_knn * 100}')
print(f'Erreur quadratique (KNN) : {mean_squared_error(y_test, y_pred_knn)}')

Erreur quadratique (Régression Linéaire) : 0.043035068198020276

Précision (Régression Logistique) : 96.76063973976687
Erreur quadratique (Régression Logistique) : 0.032393602602331255

Précision (KNN) : 99.39007861209
Erreur quadratique (KNN) : 0.006099213879100027


confusion_mat_reg_logistique = confusion_matrix(y_test, y_pred_logistique)
sns.heatmap(confusion_mat_reg_logistique, annot = True, fmt = 'd', cmap = 'Blues', xticklabels = ['Magnitude inférieure à 5', 'Magnitude supérieure à 5'], yticklabels = ['Magnitude inférieure à 5', 'Magnitude supérieure à 5'])
plt.title('Matrice de confusion Régression Logistique')
plt.xlabel('Prédictions')
plt.ylabel('Vraies valeurs')
plt.show()

confusion_mat_knn = confusion_matrix(y_test, y_pred_knn)
sns.heatmap(confusion_mat_knn, annot = True,  fmt = 'd', cmap = 'Blues', xticklabels = ['Magnitude inférieure à 5', 'Magnitude supérieure à 5'], yticklabels = ['Magnitude inférieure à 5', 'Magnitude supérieure à 5'])
plt.title('Matrice de confusion KNN')
plt.xlabel('Prédictions')
plt.ylabel('Vraies valeurs')
plt.show()


df.isnull().sum()

time               0
latitude           0
longitude          0
depth              0
mag                0
magType            0
nst                0
gap                0
dmin               0
rms                0
net                0
id                 0
updated            0
place              0
type               0
horizontalError    0
depthError         0
magError           0
magNst             0
status             0
locationSource     0
magSource          0
dtype: int64


X = df[['latitude', 'longitude', 'mag']]
y = (df['mag'] > 5).astype(int)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.7, random_state = 42)


scaler = MinMaxScaler()
X_train_norm = scaler.fit_transform(X_train)
X_test_norm = scaler.transform(X_test)


label_encoder = LabelEncoder()
y = label_encoder.fit_transform(y)
print(y)

[0 0 0 ... 1 0 0]


knn_model = KNeighborsClassifier(n_neighbors=3)
knn_model.fit(X_train_norm, y_train)
y_pred_knn = knn_model.predict(X_test_norm)


confusion_mat_knn = confusion_matrix(y_test, y_pred_knn)
sns.heatmap(confusion_mat_knn, annot = True, fmt = 'd', cmap = 'Blues', xticklabels = ['Magnitude inférieure à 5', 'Magnitude supérieure à 5'], yticklabels = ['Magnitude inférieure à 5', 'Magnitude supérieure à 5'])
plt.title('Matrice de confusion KNN')
plt.xlabel('Prédictions')
plt.ylabel('Vraies valeurs')
plt.show()


from sklearn.model_selection import GridSearchCV

param_grid = {'n_neighbors': [3, 5, 7, 9]}
grid_search = GridSearchCV(knn_model, param_grid, scoring='accuracy')
grid_search.fit(X_train_norm, y_train)

meilleurs_parametres = grid_search.best_params_
print(f'Meilleurs hyperparamètres : {meilleurs_parametres}')

meilleur_algo = grid_search.best_estimator_

y_pred_knn2 = meilleur_algo.predict(X_test_norm)
precision_knn2 = accuracy_score(y_test, y_pred_knn2)
print(f'Précision : {precision_knn2 * 100} %')

Meilleurs hyperparamètres : {'n_neighbors': 3}
Précision : 99.39007861209 %


mse = mean_squared_error(y_test, y_pred_knn)
print(f'Mean Squared Error: {mse}')

Mean Squared Error: 0.006099213879100027


print(classification_report(y_test,y_pred_knn))

              precision    recall  f1-score   support

           0       0.99      1.00      1.00     13855
           1       0.99      0.91      0.95       901

    accuracy                           0.99     14756
   macro avg       0.99      0.95      0.97     14756
weighted avg       0.99      0.99      0.99     14756


faux_positif = confusion_mat_knn[0, 1] / (confusion_mat_knn[0, 1] +
confusion_mat_knn[1, 1])
print('Taux de Faux Positifs:', faux_positif * 100)

Taux de Faux Positifs: 0.7290400972053462


param_grid = {'n_neighbors': [3, 5, 7, 9], 'metric': ['euclidean', 'manhattan', 'minkowski', 'cosinus']}

grid_search = GridSearchCV(knn_model, param_grid, scoring='accuracy')
grid_search.fit(X_train_norm, y_train)

Parametres_opti = grid_search.best_params_
print(f'Meilleurs hyperparamètres : {Parametres_opti}')

meilleur_algo = grid_search.best_estimator_

y_pred_knn3 = meilleur_algo.predict(X_test_norm)
precision_knn3 = accuracy_score(y_test, y_pred_knn3)
print(f'Précision (KNN avec meilleurs paramètres) : {precision_knn3 * 100}')

D:\anaconda3\Lib\site-packages\sklearn\model_selection\_validation.py:378: FitFailedWarning:


20 fits failed out of a total of 80.
The score on these train-test partitions for these parameters will be set to nan.
If these failures are not expected, you can try to debug them by setting error_score='raise'.

Below are more details about the failures:
--------------------------------------------------------------------------------
20 fits failed with the following error:
Traceback (most recent call last):
  File "D:\anaconda3\Lib\site-packages\sklearn\model_selection\_validation.py", line 686, in _fit_and_score
    estimator.fit(X_train, y_train, **fit_params)
  File "D:\anaconda3\Lib\site-packages\sklearn\neighbors\_classification.py", line 207, in fit
    return self._fit(X, y)
           ^^^^^^^^^^^^^^^
  File "D:\anaconda3\Lib\site-packages\sklearn\neighbors\_base.py", line 446, in _fit
    self._check_algorithm_metric()
  File "D:\anaconda3\Lib\site-packages\sklearn\neighbors\_base.py", line 381, in _check_algorithm_metric
    raise ValueError(
ValueError: Metric 'cosinus' not valid. Use sorted(sklearn.neighbors.VALID_METRICS['brute']) to get valid options. Metric can also be a callable function.


D:\anaconda3\Lib\site-packages\sklearn\model_selection\_search.py:953: UserWarning:

One or more of the test scores are non-finite: [0.99399072 0.99351616 0.99161868 0.99082804 0.99399072 0.99351629
 0.99177703 0.9900379  0.99399072 0.99351616 0.99161868 0.99082804
        nan        nan        nan        nan]

Meilleurs hyperparamètres : {'metric': 'euclidean', 'n_neighbors': 3}
Précision (KNN avec meilleurs paramètres) : 99.39007861209


if faux_positif > 0.01:
    print('L\'intégrité de vos données se dégrade, veuillez réctifier votre modèle ')
else:
    print(f'Votre modèle propose des données cohérentes, le taux de faux positif est de {faux_positif * 100} %')

Votre modèle propose des données cohérentes, le taux de faux positif est de 0.7290400972053462 %

Projet IA : Tremblements de terre 2023 mondiaux¶

Cours: Projet Data Science & Intelligence Artificielle

Professeur: Nancy CHENDEB

Date de rendu: 28 Janvier 2024

Trinôme: Nicolas RENOIR (CCC) 33% - Killian FOURNIER (CCC) 33% - Lucas LOPES (CCC) 33%

Sommaire:

1- Nettoyage et prétraitement des données

2- Analyse exploratoire des données (EDA)

3- Visualisation

4- Sélection de caractéristiques cibles

5- Sélection du modèle

6- Entraînement du modèle

7- Evaluation du modèle

8- Optimisation, maintenance et mise à jour

Chargement des données¶

1 - Nettoyage et prétraitement des données¶

2 - Analyse exploratoire des données (EDA)¶

3 - Visualisation¶

4 - Sélection de caractéristiques cibles¶

5 - Sélection du modèle¶

La régression linéaire simple¶

La régression linéaire multiple¶

La régression logistique¶

K-means¶

KNN¶

HCA¶

Évaluation des algorithmes¶

Evaluation de l'erreur quadratique et de la précision des modèles :¶

Conclusion :¶

6 - Entraînement du modèle¶

7 - Evaluation du modèle¶

8 - Optimisation, maintenance et mise à jour¶

Optimisation du modèle :¶

Maintenance du modèle :¶

Mise à jour des données :¶

	time	latitude	longitude	depth	mag	magType	nst	gap	dmin	rms	...	updated	place	type	horizontalError	depthError	magError	magNst	status	locationSource	magSource
0	2023-01-01T00:49:25.294Z	52.0999	178.5218	82.770	3.10	ml	14.0	139.0	0.8700	0.18	...	2023-03-11T22:51:52.040Z	Rat Islands, Aleutian Islands, Alaska	earthquake	8.46	21.213	0.097	14.0	reviewed	us	us
1	2023-01-01T01:41:43.755Z	7.1397	126.7380	79.194	4.50	mb	32.0	104.0	1.1520	0.47	...	2023-03-11T22:51:45.040Z	23 km ESE of Manay, Philippines	earthquake	5.51	7.445	0.083	43.0	reviewed	us	us
2	2023-01-01T03:29:31.070Z	19.1631	-66.5251	24.000	3.93	md	23.0	246.0	0.8479	0.22	...	2023-03-11T22:51:29.040Z	Puerto Rico region	earthquake	0.91	15.950	0.090	16.0	reviewed	pr	pr
3	2023-01-01T04:09:32.814Z	-4.7803	NaN	63.787	4.30	mb	17.0	187.0	0.4570	0.51	...	2023-03-11T22:51:45.040Z	99 km SSW of Pagar Alam, Indonesia	earthquake	10.25	6.579	0.238	5.0	reviewed	us	us
4	2023-01-01T04:29:13.793Z	53.3965	-166.9417	10.000	3.00	ml	19.0	190.0	0.4000	0.31	...	2023-03-11T22:51:38.040Z	59 km SSW of Unalaska, Alaska	earthquake	1.41	1.999	0.085	18.0	reviewed	us	us
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
26637	2023-12-29T03:37:19.334Z	-6.9527	154.9829	10.000	5.20	mb	72.0	60.0	3.9240	0.93	...	2023-12-29T04:05:57.040Z	89 km SW of Panguna, Papua New Guinea	earthquake	10.07	1.765	0.048	141.0	reviewed	us	us
26638	2023-12-29T04:38:54.109Z	32.3262	141.7386	10.000	5.10	mb	74.0	121.0	1.8030	0.70	...	2023-12-29T10:59:44.533Z	Izu Islands, Japan region	earthquake	9.17	1.870	0.042	187.0	reviewed	us	us
26639	2023-12-29T08:42:05.747Z	-7.2411	68.0663	10.000	5.10	mb	60.0	54.0	12.7760	0.57	...	2023-12-29T08:57:05.040Z	Chagos Archipelago region	earthquake	8.02	1.792	0.090	40.0	reviewed	us	us
26640	2023-12-29T11:02:48.679Z	-19.1602	169.0428	153.264	4.70	mb	40.0	61.0	3.7460	0.82	...	2023-12-29T11:22:46.040Z	49 km NNW of Isangel, Vanuatu	earthquake	8.52	7.433	0.081	46.0	reviewed	us	us
26641	2023-12-29T16:31:16.656Z	25.1050	96.5309	10.000	5.00	mb	53.0	64.0	4.1560	0.78	...	2023-12-29T16:45:27.040Z	92 km WSW of Myitkyina, Myanmar	earthquake	7.85	1.876	0.071	64.0	reviewed	us	us